Global search FAQs
Learn more about global search and how it works.
Global search indexing
Q. How does the indexing work?
A. IBM® OpenPages® global search server includes an indexing service. When you click Create in the Administration UI, the indexing service queries the OpenPages database for records of object types and fields that are enabled for global search. Those records are read, formatted, and then indexed by Apache Solr. The index is later used for search when the global search feature is used from the OpenPages UI.
This one-time initial indexing is called full indexing. After full indexing is complete, the global search indexer service enters an incremental indexing mode. Incremental indexing mode queries the OpenPages database once a minute for modified records, newly added records, or deleted records of object types that are enabled for global search and reindexes them to keep the global search index in sync with the OpenPages data.
Q. Does global search crawl the network-attached storage (NAS) (openpages-storage location)?
A. OpenPages uses file storage, and shares the file storage location across all instances of OpenPages applications. By default, OpenPages global search is configured with file search enabled. So yes, if file search is enabled, access to openpages-storage location must be set for global search. For more information, see The global search properties file .
Q. Are there any regularly scheduled batch jobs?
A. Yes, there is a regularly scheduled batch job that runs. The indexer (when it is running in incremental indexing mode) runs as a regularly scheduled batch job.
Q. Are the indexes and logs stored on a local drive?
A. OpenPages global search index is a proprietary format of Apache Solr, and is stored on the local hard disk drive where the global search server is installed. OpenPages global search can be installed on any local hard disk drive on the system where you installed global search. For optimal performance, it is best to install global search on a separate solid-state drive, that is, not on the same drive as the operating system.
Q. How is the index used by the product?
A. The index is proprietary to Apache Solr, is used by Apache Solr only, and is stored on the local file system where global search is installed. The index is accessed and updated anytime global search is indexing in both full and incremental mode, and the index is looked up anytime an OpenPages user is running a search using global search.
Q: How do I drop and re-create the index?
For information about dropping and re-creating the index, see Re-creating the index for global search.
Sizing and scaling
Q. The documentation mentions that global search can be run only as a single instance.
A. This is correct. A single instance of OpenPages global search server can handle all search requests from all OpenPages application servers.
Q. Can a single search server be shared by multiple environments, such as Development, Testing, Staging, or Production?
A. No. Each environment must have its own OpenPages global search server.
Q. We have two application servers and two Cognos® servers for staging and production; do we need to have two global search servers?
A. A single OpenPages global search server works with all OpenPages application servers and Cognos servers. It is strongly recommended to have a separate global search server for staging and production.
Q. Is there a possibility of running global search when using F5 load balancing in an active-active scenario?
A. Because there is only a single instance of OpenPages global search server, there is no need for F5 load balancing.
Q. How does global search impact the database server?
A. During full indexing mode, the database server is used to serve records to the global search indexer. The same is true during incremental indexing mode, but the load on the database is less because only modified records and newly added records are read. When a user uses global search and complex security rules exist for that user, those rules are processed, so the complexity of the rules determines the impact on the database server.
Q. Do we have any metrics for indexing volumes? What is the speed? Do we have any indicative query response times?
A. There are several factors influence the indexing volume and duration. The number of object types and their fields you enabled for global search and the total number of records, as well as the data size in those records in your database, all influence how long it takes to complete the initial full indexing and how large the index size on disk is. Query response time is within seconds for most searches. In some cases, if a user has complex security rules, it's possible a search can take more than several seconds.
High Availability (HA) and Disaster Recovery (DR)
Q. We have two application servers; one acts as an active administration server and the other one is a passive disaster recovery (DR) server. Do the global search servers also need to be active and passive?
A. No. Because there is only one instance of OpenPages global search server, there is no active / passive setup for it.
Q. Indexing can be lost in a DR scenario. In a DR scenario, does global search require reindexing?
A. Yes. This requires you to fully reindex for the global search feature to become available again. However, all other OpenPages functionality continues to work as normal while the global search index is being re-created.
Maintenance
Q. Are there any backup and restore operations to be performed on the global search server?
A. There are no backup and restore requirements for OpenPages global search. However, if you restore the database from a backup, the global search index is now out of sync with the OpenPages database. In this scenario, you must re-create the global search index by first disabling global search, then dropping the index, and then creating the index.
Q. Are there any index optimizations?
A. There is no need to optimize the OpenPages global search index. Apache Solr dynamically and automatically optimizes the index over time. If you perform a bulk update that impacts over 50% of your records and you have many records - hundreds of thousands of records, for example - the automatic index optimization of Apache Solr can take several days to catch up. If you suspect search performance is suffering because of a bulk update, you can force an index optimization from the Apache Solr administration page.